Addressing the quality of care at the end of life in the U.S., the 2004 NIH State of the Science Conference on Improving End-of-Life Care and the 2011 Summit on the Science of Compassion sponsored by the National Institute of Nursing Research describe an urgent need for studies that: (1) go beyond a single disease cross- sectional approach to use mixed methods in examining complex co-morbidities in longitudinal studies of established cohorts and (2) explore how and why treatment decisions are made from patient and family perspectives. Our proposal represents a rare and time-limited opportunity to address these issues in a study of end-of-life decision making in a cohort of 1337 physician graduates of The Johns Hopkins University School of Medicine ranging in age from 71 to 95 years and living throughout the United States. The aims of this mixed methods study are: (1) to assess personal, health, and functional characteristics and health service use associated with preferences for end-of-life care and planning over the course of a 15-year follow-up interval; (2) to assess how preferences and change in preferences for end-of-life care and planning are shaped by personal and professional experiences with end-of-life decision making; and, (3) to understand the experience of end-of-life decision making from the point of view of family members among participants who died, in relation to preferences for end-of-life care and planning expressed prospectively and to treatment actually received in the last year of life. We employ mixed methods, combining the latent transition model (quantitative methods) with narrative and thematic analyses (qualitative methods). For Aim 1, we employ the latent class transition model examining personal characteristics associated with preferences for end-of-life care and planning over time and association with health services use. In Aim 2, for a sample of participants we will carry out semi-structured interviews to elicit the respondent's point of view about goals of treatment and planning for end-of-life care, and rationale behind preferences over the course of 15 years (e.g., from fitting a category of preferences in which most potentially life-sustaining treatments are desired to one in which few are). For Aim 3, we will use semi- structured interview methods to elicit narratives from family members about experiences with end-of-life decision-making among participants who have died. By allowing individuals to describe decision making through narrative, we learn how people think about and define their own goals for end-of-life care differently than clinicians and researchers. The mixed methods study provides an unprecedented life span perspective on end-of-life preferences and planning as participants pass through medical, functional, and social transitions. We will learn in a way not possible in other studies how older adults would chose to have decisions made and with whom, and why. The study can change the public discourse on end-of-life care and planning.